Fanni Csincsak

I ship AI where it has to actually work — fintech, SaaS, and regulated enterprise — and I can prove it survived governance.

Twelve years across AI, fintech, and enterprise ITSM. Right now I lead AI and workflow product at T-Systems International (Deutsche Telekom). Before that, scaleups and early-stage: a fintech platform to 24,000 users, an AI legal-tech product from zero.

What that's looked like:
- An AI change assistant that cut request time from 15 minutes to 7, and lifted compliance from 72% to 89%. We were asked to cover 2,000 templates. The usage data said most were barely touched, so we built it for the ones people actually use.
- An AI governance framework covering 40 AI use cases and roughly 509 hours a month — with DORA risk assessment at intake. About one in ten submissions turned out not to be AI at all. Catching that early is the job.
- An AI/AIOps integration running live across six enterprise customers.

Before T-Systems: a fintech platform that reached 24,000 users in under six months, and an AI legal-tech product at AltaML where 20+ user interviews killed the destination app we were about to build — and turned it into a browser extension instead. It shipped, and it anchored $1M+ in follow-on funding.

The pattern doesn't change. The brief says one thing. The evidence says another. I go with the evidence, and I bring the numbers.

Most people who call themselves AI product managers have shipped a prototype. I've taken AI into a regulated European enterprise and made it pass audit — which, with the EU AI Act phasing in through 2026 and 2027, is turning out to be the harder half.

What I work on: AI product strategy · AI governance & EU AI Act readiness · applied AI and LLM features in enterprise workflow · product discovery and requirements · roadmap prioritisation · ServiceNow (NowAssist, Predictive Intelligence, Incident, Change, Problem, Knowledge) · fintech funnels, activation and retention · GDPR & DORA · Scrum, Kanban, BDD · PostHog, GA4, Amplitude.

Budapest-based, EU work-eligible, remote-first. Hours overlap London, Berlin, Amsterdam, Copenhagen.

Case studies: fannicsincsak.com